JPMorgan's recent trading loss has prompted many on Capitol Hill to call for
stricter and more immediate enforcement of the Dodd-Frank Wall Street Reform and
Consumer Protection Act. Sadly, more stringent enforcement of Dodd-Frank will only
make markets more volatile and ensure that more JPMorgan-like trading losses are
incurred moving forward. To understand why Dodd-Frank will pervert incentives and
cause market disorder, one need only look at the tools the legislation utilizes to gauge the
risk of and collateral needed by "Major Swap Participants" (essentially, institutions that
have large derivatives holdings): risk-weighting and value at risk (VAR).
First, lets look at risk-weighting. To determine if an entity qualifies as a closely
supervised Major Swap Participant, (MSP) Dodd-Frank assigns risk-weighted
multiples to its derivatives positions to measure exposure. Essentially, this calculates a
position's risk by multiplying its face value by a preconceived risk matrix with values
for the given position type. For example, a credit swap with a maturity of two years
gets a risk-weighting of 10%. This means that a derivative contract with a $1 million face
value would be risk-weighted at $100,000 -- this number is arrived at by multiplying the
face value of $1 million by the risk-weighting multiple of 10%.
This perverts markets by incentivizing traders to pile into positions with low risk-weightings. The problem is, once too many folks pile into a "riskless" position, it is no
longer riskless. European banks and their adaptation of Basel II's capital rules provide
an example. Essentially, Basel II risk-weighted banks' assets to determine how much
capital they had to hold. Sovereign debt was afforded a zero risk-weighting, so banks
chased yield and piled into sovereign debt. The upshot of such distortion is the
destructive interconnection of today's European economy, as banks are dependent on the
fiscal health or the sovereign and the sovereign is reliant on the continued survival of the
banks.
Interestingly, under Dodd-Frank's risk-weighting system, interest rate swaps with
a maturity of less than one year are afforded a zero risk-weighting. This includes
positions on LIBOR. Remember -- LIBOR spreads spiked 431 basis points (4.31%) in a
single day, rising from 257 basis points to 688 basis points on September 30, 2008.
Emulating Europe's example on risk-weighting is the equivalent of following Greece's
lead on fiscal management.
The next Dodd-Frank risk barometer utilized for Major Swap Counterparties, is
VAR, or value at risk. This is used to determine how much cash counterparties should
set aside for their derivatives contracts. The problem with VAR is that it uses historical
market data to predict future risk. This backward-looking bias featured prominently in
the American International Group (AIG) debacle -- housing prices had never experienced a
deep and uniform fall, so VAR models indicated that bets on mortgages were safe.
Again, this distorts markets, and incentivizes market actors to pile into
historically "riskless" assets with high yield and low VAR. Market failures are not
typically caused by assets widely acknowledged to be risky at inception, which would be
most sensitive to VAR measurement. JPMorgan was not betting on Greek bonds, it was
betting on highly rated investment grade corporate debt.
These tools -- risk weighting and VAR -- will be ineffective, and engender more, not
less, risk. They will also hurt liquidity by promulgating a string of seemingly arbitrary
capital requirements for basic derivatives transactions. An easier and more effective
alternative should be devised by analyzing what most large, market-roiling, trades have
in common.
Market failures typically involve a counterparty with a highly concentrated
position, comprising a significant and unusual proportion of the respective marketplace
or index's total value. The product underlying it is usually riskless or relatively safe.
This "safe" product is then made "unsafe" by a combination of the unusual and
concentrated position of the given counterparty and a change in macro conditions.
As an example, JPMorgan was insuring an index of investment grade corporate
bonds. This insurance cost about 1% of face value per year -- essentially the same as the
cost to insure against the default of Germany -- so it's fair to say this was relatively safe.
JPM held about $100 billion of wagers in the index, and according to clearinghouse data,
the total face value of positions on the index was about $785 billion, giving JPM nearly a
13% share of the index's total face value. Granted, the $785 billion does not include
trades not put through a clearinghouse, but it's apparent that they had a highly
concentrated position. Gyrations only began once the macro woes of Europe caused
credit spreads to widen on the corporate debt.
The impact that a concentrated position holder can have on the risk of an asset is
not confined to recent financial history or complex debt derivatives. The Hunt brothers'
attempt to corner silver markets in 1980 perhaps best exemplifies the power of large and
concentrated positions to roil markets and redefine price-range movements. The Hunt
Brothers began buying silver en masse in 1979, and by January of 1980 they held or
controlled over half of the world's silver supply. The average price of silver had never
been more than $6/ounce over the course of a year, but prices spiked to over $50/ounce
in January of 1980 due to the Hunt brothers' continued accumulation. This is an extreme
example, but the point is, no risk-weighting or VAR system could have predicted such
stark price movements using historical data.
This is not a reflection of the competence of regulators. It is a reflection of the
fact that large losses are most often caused by unpredictable price movements that are
impossible to predict using historical pricing data. Instead of attempting the impossible,
and impeding liquidity in the process, regulators should utilize the information that they
do have. One good thing about Dodd-Frank is its mandate that MSPs disclose their
positions in regulatory filings. Regulators should aggregate this information and identify
the relative concentrations of counterparties in derivatives positions.
If regulators observe a certain position rising to a heightened level in terms of
concentration for a given index or market, they can flag the position and assess current
capital requirements. If the concentration level has a precedent, they should look to the
volatility of the index or market when the concentration level was commensurately high
in the past, to assess the capital currently needed. This would be far more relevant to
determining the real "value at risk" than simply inputting the last five years of historical
data or stress-testing. If the concentration is unprecedented, regulators can assess
whether heightened capital requirements should be imposed to compensate for the risk
engendered by such uncharted territory.
The danger of concentration is not limited to single counterparties. For example,
what if two or three counterparties are dominating an index or marketplace, collectively,
in an unprecedented manner? This problem is also not intractable. As long as regulators
have the aggregated trade data, they can monitor such collective concentrations. One
potential solution could be a Herfindahl-Hirschman Index (HHI) model, frequently used
in antitrust cases to measure market concentration amongst firms to assess competitive
balance. Essentially, the HHI is higher in less competitive and more concentrated
industries -- if it was employed to measure index or marketplace concentrations in a
similar fashion, an unusually high HHI figure could flag latent danger.
JPMorgan's trading loss does raise questions, but we shouldn't be so desperate
for answers that we grasp aimlessly. We should instead add historical context to the JPM
fiasco, and utilize its commonality to better screen for similar potential scenarios moving
forward.
Matthew Schoenfeld can be reached at mschoenfeld@jd12.law.harvard.edu